- From: Arthur Ryman <arthur.ryman@gmail.com>
- Date: Wed, 24 Jun 2015 07:33:18 -0400
- To: Holger Knublauch <holger@topquadrant.com>
- Cc: public-data-shapes-wg <public-data-shapes-wg@w3.org>
Holger, I understand that this may be seen as a subtle distinction. I believe we agree that we should at least be consistent with the use of "Class" as in sh:scopeClass and sh:valueClass. You believe that most new SHACL users will want to use sh:valueClass. We can even recommend its use. We can describe sh:valueType as being defined for backward compatibility with oslc:valueType. OSLC users will feel comfortable with this term. -- Arthur On Tue, Jun 23, 2015 at 10:15 PM, Holger Knublauch <holger@topquadrant.com> wrote: > On 6/24/2015 10:50, Arthur Ryman wrote: >> >> +1 >> >> However I suggest a name change. >> >> Use sh:valueClass in Proposal 2. This consistent with sh:scopeClass in >> Proposal 4. >> Use sh:valueType in Proposal 3. > > > My concern is how to explain this difference to users. This distinction is > very subtle and many people will make mistakes unless we select very clear > names. With valueClass and valueType I would myself be confused, because the > choice feels arbitrary to me. I could live well with sh:valueClass replacing > the current sh:valueType (also because it aligns well with sh:scopeClass). > But then I think we need something more distinct for the case of direct > rdf:type without subclasses. We can hopefully solve this quickly during the > next meeting. > > Holger > >
Received on Wednesday, 24 June 2015 11:33:46 UTC